Impact of screening for large-for-gestational-age fetuses on maternal and neonatal outcomes: a prospective observational study

Introduction

Large-for-gestational-age (LGA) fetuses are defined by a fetal weight exceeding the 90th percentile [1], 2] for a given term, or 4,000 g [3], [4], [5] at birth. Delivery of an LGA foetus can lead to a number of obstetric complications. Some of these can be serious [6], 7]. Fetal and neonatal morbidity and mortality increase from 4,000 g [8], and especially above 4,500 g [9], with shoulder dystocia being the most frequent cause. Mothers are also at risk, facing higher rates of postpartum hemorrhage and perineal tears [6], 7]. The management of LGA fetus is a daily problem in delivery rooms, with its management often unclear and its definition disputed.

In our hospital, protocols for managing LGA fetuses are based on screening. A third-trimester ultrasound is systematically performed around 32 weeks of pregnancy, as is common in France, to detect fetal pathologies (abnormal growth, malformations, etc.). If the estimated fetal weight is excessive (above the 90th percentile according to the hospital’s curves), a growth ultrasound is scheduled around 36 weeks. Due to the imprecision of screening, many fetuses falsely identified as LGA may be unnecessarily managed with labor induction. Even for accurately identified LGA fetuses, this can induce prematurity, leading to increased morbidity and mortality up to one year, even after 38 weeks of pregnancy [10], [11], [12]. The idea that the fetus may be large-for-gestational-age could also have an impact on the method of delivery (vaginal delivery with or without an instrument, or Cesarean section). Cesarean sections have an impact on maternal [13], [14], [15] and neonatal [16], [17], [18] morbidity and mortality, and potentially on the long-term outcome for the child. We therefore need to study the impact of our screening on the route of delivery.

The effect of ultrasound macrosomia on birth parameters has been studied recently. However, the studies used a threshold of 4,000 g to define macrosomia, which is very high for the condition to which it was applied and is therefore rare and indicative of severe situations [19]. In other cases, studies have compared the result of a third trimester ultrasound scan with no ultrasound scan [20]. The effect of an organised screening protocol with a systematic ultrasound scan in the third trimester and an additional scan one month later, even in cases of mild macrosomia (>90th percentile), is therefore not really known. This allows for the effect of false positives: changing how eutrophic neonates mistakenly identified as LGA are managed. Finally, using a fixed threshold may be questionable. The definition of a macrosomia threshold in relation to maternal characteristics needs to be investigated.

The aim of this study is to investigate the consequences of a systematic screening protocol for fetal macrosomia, which is in use in several institutions in the Champagne-Ardenne region (France). This will help to determine whether screening for LGA fetuses can prevent risks by anticipating them, and whether it has a negative impact on false-positive fetuses.

Materials and methods

Results

A total of 2,217 women were included in the cohort follow-up. Population characteristics are presented in Table 1 (univariate analysis). The study flowchart is shown in Figure 1.

Figure 1:

Study flowchart.

Discussion

This study shows that simply knowing that a fetus is LGA could significantly increase cesarean and instrumental vaginal delivery rates without improving maternal or neonatal outcomes. Notably, our study is the first to demonstrate a significant difference in the rate of instrumental deliveries.

Earlier studies [19], 20], [25], [26], [27], [28], [29] have identified some of these consequences. First, the definition of macrosomia used, the 90th percentile according to adjusted AUDIPOG [22] curves, is not an absolute weight. It is based on maternal characteristics (height, weight, parity) and fetal characteristics (sex, gestational age). It seems logical that the same neonatal weight would not have the same impact on a patient depending on her height and weight. Authors such as Ye et al. [30], have already highlighted the value of these curves in emphasing population-specific adjustments for LGA definitions. In addition, this improves the external validity of the results, as this definition is the one used in practice in many French maternity units. Finally, unlike this definition, an absolute weight is not adjusted for gestational age.

Other authors [25], [26], [27] have used a definition based on a higher weight (4,000 g or the 97th percentile). This could be considered “severe”. This means fewer events and these decisions could significantly improve the negative predictive value of screening. Professionals faced with these higher weight estimates are likely to be even more concerned about neonatal or maternal complications. The work using a broader definition of macrosomia based on the 90th percentile is interesting because it also considers the uncertainty of screening and the impact of macrosomia considered less critical in practice. To go further, the ultrasound definition of LGA fetuses was the same (90th percentile of antenatal curves), but the definition of LGA newborns is more stringent in these authors. Therefore, the detection rate of LGA fetuses is logically lower than ours. However, the impact of this detection rate is difficult to assess, as it is still low despite improvements.

Because in our clinical setting teams systematically request these measurements when there is a suspicion of macrosomia, studies [20] comparing weight estimation with no estimation do not seem relevant in practice. Therefore, the non-estimation of the weight seems to be purely experimental and could even lead to more tension for professionals, who might otherwise assume that the fetus is LGA on the basis of their clinical examination or other ultrasounds.

Screening for LGA fetuses significantly raised cesarean rates in neonates mistakenly considered LGA (false positives), without significant differences in neonatal or maternal outcomes. This is consistent with the existing literature [28], 29] but contrasts with our findings of increased instrumental extraction rates.

Newborns in the false-positive group weighed approximately 400 g more than controls, probably due to the margin of error of ultrasound. This overestimation does not justify higher cesarean delivery rates, as average weights (∼3,600 g) typically do not cause feto-pelvic disproportion. Despite the change in the mode of delivery, no improvement in neonatal or maternal complications was observed. However, the rarity of these events may mean that our study was underpowered to detect significant differences.

Interestingly, the reasons for cesarean delivery remained consistent across groups, contradicting expectations of increased cesarean delivery for suspected LGA fetuses, as previously described [25]. This suggests that the impact of suspected LGA status is broad and not limited to labor interpretation.

The immediate focus of our study limits understanding of long-term effects, such as postpartum depression or poorer birth experiences due to medical intervention, as previously described [31], 32]. The over-medicalization of such screening contributes to unnecessary complications and increased health costs. This is part of the broader problem of “too much, too soon” medicalization, which leads to less appropriate situations for women and higher health costs [33].

There was also a difference in the rate of induction between the groups in our study. This difference was expected. There was significantly more induction of labor in the groups with suspected LGA fetuses. Labor induction is considered by some groups to be a risk factor for cesarean section and could explain these results. However, these studies are questionable because of the control group. In fact, the situation that should be compared to induction of labor is not spontaneous labor, but rather a wait-and-see approach because of the possibility of adverse obstetric events that would not occur in the context of artificial induction of labor. On the other hand, studies comparing induced labor with a wait-and-see approach seem to indicate a lower rate of cesarean section and should therefore have protected the LGA groups in our study, which is the opposite result obtained.

Induction rates varied as expected, with more inductions in the suspected LGA groups. However, studies comparing induction with a wait-and-see approach indicate lower cesarean delivery rates [34]. This does not explain our findings.

Our detection rate for suspected LGA birth was higher than other studies [25], 26], possibly because of different curves. However, all curves to date have so far shown poor performance in the detection of fetal mascrosomia [35]. This difference may also be due to the 90th percentile threshold, which we feel is more in line with our daily practice.

Despite this improvement, the detection rate remains unsatisfactory. To further improve the practice of LGA, it seems essential to improve our diagnostic performance. Some authors have proposed the use of Doppler markers [36] and maternal characteristics with biomarkers [37] for this purpose. This would reduce the number of unnecessary interventions in normal weight newborns who have been misclassified as AGL. However, it does not solve the problem of over-medicalization of LGA newborns, as known newborns in our study did not have better outcomes than unknown newborns.

Staff at our inclusion centers had similar training, but LGA protocols in France vary widely, limiting the generalizability of our results.

The mode of delivery was chosen as the main criterion. This choice was made because it is now known that the mode of delivery has an impact on maternal [13], [14], [15] and neonatal [16], [17], [18] morbidity. In addition, because the other criteria are much less common, it would have been difficult to highlight any differences.

A subgroup analysis for severe macrosomia (>97th percentile) was planned, but the sample size was too small to be of real interest.

This prospective, multicentrer study ran for one year, systematically enrolling eligible women and defining a uniform cohort. Excluding diabetic women brought the cohort closer to real-life scenarios, as more than 20 % of our subjects had diabetes. However, the observational design allowed weaker control of screening bias and lacked centralized review of examinations, which could influence healthcare professionals’ decisions – central to our research interest.

Given the limited size of our population, it is difficult to draw firm conclusions from our results regarding the management of LGA neonates, particularly regarding decisions about cesarean vs. instrumental delivery. In addition, these decisions are in part subjective and are influenced by the beliefs of the health care professionals (both physicians and midwives) [38]. However, as this study is consistent with the findings of several others, it may now be relevant to encourage professionals to consider that the anticipation of LGA fetuses does not necessarily improve their outcomes. Being aware that their reasoning may be biased by the classification of the fetus as LGA may have an impact on decision making in such situations.

Compared to fetuses with more severe weight estimates, these considerations may be particularly important for fetuses with weights near the threshold. Fetuses with lower weights would logically experience fewer complications related to macrosomia. However, they are also subject to cognitive bias due to a threshold effect. These notions, in combination with the limited performance of ultrasound, may have implications for a re-evaluation of medical intervention decisions.

This study contributes to evidence-based decision making for late pregnancy ultrasound protocols by highlighting the limitations and potential biases associated with estimating fetal weight and predicting LGA outcomes. It emphasises the importance of cautious interpretation of ultrasound findings, particularly in cases where estimated fetal weights are close to diagnostic thresholds, to avoid unnecessary medical interventions driven by cognitive biases or overestimation of risk.

Although this study does not directly address the psychological impact, decisions about ultrasound protocols should take into account the potential emotional distress of expectant mothers. Being told that they are carrying an LGA fetus may increase anxiety and negatively affect their perception of childbirth. These protocols should be reserved for situations where patient management truly requires such assessments, particularly as macrosomia alone does not automatically warrant induction of labour or other interventions.

Finally, it is important to recognise that screening ‘for the sake of reassurance’ could paradoxically lead to increased stress for both patients and health care teams, undermining the intended benefits of such protocols.

Conclusions

Data from this study suggest that large-for-gestational-age screening may be associated with an increase in the rate of cesarean section for large-for-gestational-age newborns and for false-positive newborns. This would also affect instrumental delivery rates, which are higher in these newborns. In any case, it has not been shown that screening leads to a reduction in neonatal and maternal complications. This study could be another argument for discussing systematic implementation of end of pregnancy growth screening: “side effects” should perhaps be taken into account. However, these results must be weighed against the similarity of protocols and professional training in the three hospitals. These protocols may not be applicable to all maternity units.

This may be due to professional anxiety regarding the management of LGA neonates. The development of in-service training in the emergency management of such situations therefore seems essential. It would be interesting to evaluate the impact of such training on reducing maternal-fetal morbidity associated with LGA fetuses and their screening.